Database Infrastructure for Visual Information Systems

Clement Leung
Victoria University of Technology, Australia

Current information systems have evolved through several decades and have inherited many constraints prevalent in earlier generations. Next generation information systems will have a high visual content, and there will be a shift in emphasis from a paradigm of pre-dominantly alphanumeric data processing to one of visual information processing. A Visual Information System represents a radical departure from the largely text-in/text-out paradigm of conventional information systems. It will not only substantially enhance the value and usability of existing information, but will also open up a new horizon of previously untapped information sources which will greatly increase the effectiveness of an organization. However, the database characteristics and storage structures required to support a visual information system are different from those for a conventional information system. In this talk, the indexing of visual information will be examined, and mechanisms for raising the efficiency of the indexing process and enhancing retrieval effectiveness will be presented. The logical and physical structures for optimizing visual information search and retrieval will also be discussed.